Geoffrey J. Ashwell*, Rakesh Ranjan, Anne J. Whittam and Daniel S. Gandolfo
The Nanomaterials Group, Centre for Photonics and Optical Engineering, Cranfield University, Cranfield, UK MK43 0AL, g.j.ashwell@cranfield.ac.uk
First published on UnassignedUnassigned22nd December 1999
Langmuir–Blodgett (LB) films of (E)-4-[(N-octadecyl-5,6,7,8-tetrahydro-5-isoquinolylidene)methyl]-N,N-dialkylaniline octadecyl sulfate, where the donor group is dimethylamino (1a), diethylamino (1b), dibutylamino (1c) and dihexylamino (1d), have optimum susceptibilities when the dyes are co-deposited in a 1∶1 mole ratio with octadecanoic acid. Alternate-layer structures of the mixed films, interleaved with poly(tert-butyl methacrylate), are non-centrosymmetric and the second-harmonic intensity increases with thickness as to N>100 bilayers. Furthermore, when the dialkylamino group is sufficiently hydrophobic, i.e. for 1c and 1d, the dyes form non-centrosymmetric Z-type structures without interleaving spacer layers. The optimum susceptibilities, chromophore tilt angles, thicknesses and refractive indices of the interleaved and non-interleaved mixed films are as follows: alternate-layer (1b),
= 67 pm V−1 at 1.064 µm for φ = 37°, l = 4.08 nm bilayer−1, nω = 1.50 and n2ω = 1.58; Z-type (1c),
= 76 pm V−1 for φ = 33°, l = 3.15 nm layer−1, nω = 1.52 and n2ω = 1.58. The moderately high susceptibilities arise from an optimised transparency/efficiency trade-off, the films in this series being transparent at the fundamental wavelength (1.064 µm) and having a very slight absorbance of 5 × 10−4 per active layer at 532 nm.
When both component layers are SHG-active and the interfacial interactions are sterically unhindered,2 the second-order properties are disadvantaged by the fact that the intralayer and interlayer dipoles are opposed. Passive spacers may be preferable10–14 but they have the drawback of diluting the optically nonlinear component in alternate-layer films. Thus, the interleaving layer should be carefully chosen and, for this purpose, poly(tert-butyl methacrylate) is an ideal candidate.15 It readily deposits on the downstroke and is, therefore, compatible with the preferred transfer of the active dye layer on the upstroke. Furthermore, it is transparent throughout the visible and near infrared regions of the spectrum and has a thickness of only 1.03 nm per layer. Penner et al.16,17 have reported low-loss waveguide structures, with an optical attenuation of 1 to 2 dB cm−1, for alternate-layer structures of polymeric dyes and poly(tert-butyl methacrylate), and have demonstrated efficient phase-matched blue light generation with a normalised conversion efficiency of 150% W−1 cm−2 for 819 nm radiation.
In this work, we report alternate-layer structures of a cationic hemicyanine dye interleaved with poly(tert-butyl methacrylate) and demonstrate a quadratic SHG dependence, , to more than 100 bilayers. There have been several previous studies on interleaved films of hemicyanine derivatives, both monomeric11,12 and polymeric,13 but most have failed to show the theoretically expected SHG enhancement to thicknesses of more than a few bilayers.18,19 This suggests a structurally disordered arrangement and recently, Han et al.20 have reported a susceptibility enhancement from 16 pm V−1, when freshly deposited, to 50 pm V−1 when poled by a field of 1.3 MV m−1. The hemicyanine derivatives are also highly coloured and thus, combined with the difficulty of controlling the long-range structural order, they are not suitable for nonlinear optical applications.
This paper concerns a subtle modification to the molecular structure of the cationic dye compared with the previously reported hemicyanine,11–13,18–20 the suitability of its counterion, and co-deposition with octadecanoic acid to optimise local field effects and absorption in the vicinity of the second-harmonic wavelength. LB films of the modified dyes (1a–d) show improved SHG when the anion is amphiphilic, i.e. octadecyl sulfate rather than iodide, and the intensity is enhanced when the dye is diluted in a 1∶1 ratio with octadecanoic acid. The second-harmonic intensity from alternate-layer films of the dye and poly(tert-butyl methacrylate) increases quadratically with the number of active layers and, furthermore, a similar dependence has been realised to in excess of 200 Z-type layers for analogues with hydrophobically substituted donor groups.
The diethylamino to dihexylamino analogues, 1b to 1d, were obtained in a similar manner by the reaction of N-octadecyl-5,6,7,8-tetrahydroisoquinolinium iodide and the appropriate 4-(N,N-dialkylamino)benzaldehyde, the NMR and elemental data being satisfactory in each case. Salts with amphiphilic anions were obtained by metathesis of the cationic dye and sodium octadecyl sulfate at the air/water interface of the LB trough (see below). The water soluble ions, Na+ and I−, dissolve into the aqueous subphase and are not incorporated into the film.
![]() | ||
Fig. 1 Surface pressure versus area isotherm of 1b. |
Mixed films of the dye and octadecanoic acid were obtained in a similar manner, for example, by co-spreading the three components (dye∶Na+ODS−∶ODA) in a ratio of 1∶1∶n where 1 ≤ n ≤ 5. The general shape of the isotherm remains the same, when the dye is mixed with octadecanoic acid, but the combined area is progressively shifted by ca. 19 Å2 per ODA molecule for mole ratios of 1∶1 to 1∶5. Monolayer films were deposited onto hydrophilically treated glass substrates, on the upstroke, at a rate of 80 μm s−1 and at optimum surface pressures of 40 mN m−1 for 1a and 32 mN m−1 for 1b–d.
The inactive spacer, poly(tert-butyl methacrylate), used in the fabrication of the alternate-layer structures, was spread from dilute chloroform solution onto the water subphase of the second compartment of the trough. Interleaved films were then obtained by cycling a hydrophilically treated glass substrate (for SHG) or a silicon wafer (for reflectance studies), from below the surface, to deposit the mixed film (dye∶Na+ODS−∶ODA = 1∶1∶1) on the first upstroke, as above, and poly(tert-butyl methacrylate) on the subsequent downstroke at 10–12 mN m−1. Multilayer films were fabricated by repeating the process with a deposition rate of 80 μm s−1 in both directions. In contrast, multilayer Z-type structures of 1c and 1d were obtained on the upstroke by cycling the substrate via the air/water interface of the second compartment, in this case, with no floating layer.
![]() | ||
Fig. 2 Typical dependence of the second-harmonic intensity on the angle of incidence of the laser beam, relative to the LB film, for the dyes in this series. The data correspond to a Z-type film of 1d. |
Dye | λmax/nm | Amax/layer−1 | φ/° | |
---|---|---|---|---|
1a | 417 | 3 × 10−3 | 70 | 32 |
1b | 425 | 5 × 10−3 | 120 | 30 |
1c | 425 | 5 × 10−3 | 100 | 31 |
1d | 415 | 4 × 10−3 | 50 | 30 |
The linear and nonlinear optical properties are also dependent upon the film composition and optimum SHG arises when the dyes are mixed in a 1∶1 ratio with octadecanoic acid. The behaviour is similar in each case and, therefore, we report the SHG dependence of the diethylamino analogue which is representative of the series. Films of 1b exhibit absorption maxima at 425 nm for the pure dye, 450 nm for the 1∶1 mixed film and 480 nm for dye∶ODA ratios of 1∶2 to 1∶5. The SHG polarisation dependence is independent of composition and, using the method of Kajikawa et al.,21 corresponds to a chromophore tilt angle of 30 ± 1° relative to the substrate normal. In contrast, the susceptibility is strongly dependent upon the ratio of dye to octadecanoic acid with an optimum value of = 145 pm V−1 for the 1∶1 mixed film (cf. 120 pm V−1 for the pure dye). The enhancement is unlikely to be associated with resonance effects because the residual absorbance at the harmonic wavelength, albeit very small, is almost identical for each of the films of 1b. Therefore, it probably relates to improved non-centrosymmetric ordering and, as reported by Hayden22 and McGilp et al.,23 to changes in the local field effects upon dilution.
The films are almost transparent at 532 nm and thus, it may be assumed that Kleinman's symmetry is valid and that the susceptibility components are limited to and
. Furthermore, as the hyperpolarisability is probably dominated by the component along the molecular charge transfer axis, the relation between χ(2) and β is as shown in eqn. (1) and (2).
![]() | (1) |
![]() | (2) |
N is the number of molecules per unit volume, fω and f2ω are local field correction factors at ω and 2ω respectively, and f = (n2 + 2)/3 where n is the refractive index at the corresponding wavelength. The hyperpolarisability is suppressed in films with a high surface density of dye and, for 1b, saturates to a constant value of ca. 1.4 × 10−37 m4 V−1 (330 × 10−30 esu) when diluted with octadecanoic acid (Fig. 3). The apparent variation is attributed to changes in the local field correction factors and, interestingly, both β and λmax remain constant for dye∶ODA ratios of 1∶2 and above. However, as optimum SHG is realised for a dilution of 1∶1, the principal characterisation is limited to films of this composition and the following discussion only concerns this ratio.
![]() | ||
Fig. 3 Variation of the second-order susceptibility (×) and molecular hyperpolarisability (filled circles) with the surface area occupied by 1b and associated ODA molecules in the mixed LB monolayers. The nonlinear optical coefficients were evaluated using a monolayer thickness of 3.1 nm and a weighted average of the refractive indices of the dye (nω = 1.51, n2ω = 1.60) and octadecanoic acid (nω = 1.53, n2ω = 1.54) as the ratio of dye∶ODA is altered from 1∶0 to 1∶5. |
The thickness and refractive indices of an alternate-layer mixed film of the diethylamino analogue (1b) were derived from an analysis of the wavelength dependent reflection data for a 32 bilayer film (Fig. 4; Table 2). A mean thickness of 130.8 nm corresponds to 4.08 nm bilayer−1 and, from studies on non-interleaved films of poly(tert-butyl methacrylate), conforms to ca. 3.05 nm layer−1 for the SHG-active dye and ca. 1.03 nm layer−1 for the spacer. The corresponding thickness from alternate-layer films of the dibutylamino analogue (1c) is 4.13 nm bilayer−1 and conforms to 3.10 nm layer−1 for the dye compared with 3.15 nm layer−1 obtained from the reflectance data of a non-interleaved Z-type film of 1c. Thus, for these analogues, a minimal difference in the layer thickness suggests that the dialkylamino groups are expanded in the horizontal rather than the vertical direction. Sufficient space is made available, at the other end, by a side-by-side packing arrangement of the three octadecyl tails of the cationic dye, the amphiphilic anion and the octadecanoic acid.
![]() | ||
Fig. 4 Reflectance of polarised light from 32 bilayers of the 1∶1 mixed film of 1b and octadecanoic acid interleaved with poly(tert-butyl methacrylate) on a silicon wafer: (a) λ = 543.5 nm; (b) 594.1 nm; (c) 611.9 nm; (d) 632.8 nm. The curves represent the theoretical fits using the data in Table 2. |
Coefficient | Wavelength/nm | |||
---|---|---|---|---|
543.5 | 594.1 | 611.9 | 632.8 | |
nx | 1.565 | 1.547 | 1.540 | 1.538 |
ny | 1.557 | 1.526 | 1.520 | 1.515 |
nz | 1.583 | 1.544 | 1.530 | 1.526 |
niso | 1.568 | 1.539 | 1.530 | 1.526 |
kx | 7.0 × 10−3 | 1.0 × 10−4 | 0 | 0 |
ky | 0.3 × 10−3 | 1.7 × 10−4 | 0 | 0 |
kz | 1.5 × 10−2 | 5.0 × 10−4 | 0 | 0 |
l/nm | 130.5 | 130.4 | 130.8 | 130.6 |
l/nm bilayer−1 | 4.078 | 4.075 | 4.088 | 4.081 |
The refractive indices, obtained from the analysis of the reflectance data of the alternate-layer film of 1b, increase with decreasing wavelength and their dispersion may be represented by eqn. (3), the Sellmeir equation (Fig. 5).
![]() | ||
Fig. 5 Variation of the isotropic refractive index with wavelength of a 1∶1 mixed film of 1b and octadecanoic acid interleaved with poly(tert-butyl methacrylate). The points represent the experimental data from Table 2 and the solid line is the theoretical fit using the Sellmeir equation. |
![]() | (3) |
where A is a constant, s is the oscillator strength and λ0 is the resonance wavelength of the absorbance band. The refractive indices at the fundamental and harmonic frequencies, obtained by substitution of the experimental data into eqn. (3), are 1.495 and 1.581 respectively. The difference (n2ω − nω = 0.086) is smaller than reported26 for LiNbO3 and substitution into the following equation provides a coherence length of ca. 3 μm at normal incidence for the wavelength described in eqn. (4).
![]() | (4) |
The data obtained for the dibutylamino analogue (1c) are summarised in Table 3 for alternate-layer and Z-type films. The thickness and refractive indices of the alternate-layer structure are similar to those obtained for 1b.
Dye | Dye∶ODA | Thickness/nm layer−1 | nω | n2ω | Film type |
---|---|---|---|---|---|
aCorresponds to the repeating bilayer structure of the alternate-layer film interleaved with poly(tert-butyl methacrylate). | |||||
1b | 1∶1 | 4.08a | 1.50 | 1.58 | interleaved |
1c | 1∶1 | 4.13a | 1.51 | 1.55 | interleaved |
1c | 1∶1 | 3.15 | 1.52 | 1.58 | Z-type |
1c | 1∶0 | 3.16 | 1.52 | 1.59 | Z-type |
The apparent quadratic SHG dependence of the 1∶1 mixed films of 1a and 1b is shown in Fig. 6 but, if viewed as /N2vs.N, there are discrepancies in the normalised intensities. These probably result from variable deposition conditions. The multilayer structure was fabricated in several stages, throughout a period of two weeks, this tedious and unnecessary procedure resulting from our enforced requirement to monitor the linear and nonlinear optical behaviour throughout the entire deposition process. The diethylamino analogue (1b) exhibits the strongest SHG but the normalised intensity decreases by a factor of two throughout the early stages of deposition. However, it is constant for N > 55 bilayers and the deviation is also accompanied by a slight shift of the peak wavelength, from an initial value of 455 nm to 447 nm for thicknesses greater than 60 bilayers. Furthermore, the SHG polarisation dependence, I2ω(p → p)/I2ω(s → p), indicates that the chromophores modify their tilt from ca. 30° in the monolayer to 37° in the thick alternate-layer film, both values being relative to the normal to the substrate.
![]() | ||
Fig. 6 Variation of the square root of the second-harmonic intensity with the number of bilayers for 1∶1 mixed films interleaved with poly(tert-butyl methacrylate): 1a (×); 1b (solid circles). |
The second-order susceptibility decreases from
≈ 110 pm V−1 for the first bilayer to a steady state of ca. 67 pm V−1 for structures in excess of 55 bilayers. Despite this, the properties are extremely encouraging and optimisation of the deposition conditions should result in improved second-order coefficients. The normalised SHG,
, from the thick multilayer film is significantly greater than the signal obtained from LB films of the extensively studied hemicyanine, (E)-N-docosyl-4-[2-(4-dimethylaminophenyl)ethenyl]pyridinium bromide, first reported by Girling and co-workers.27 In addition, whereas the hemicyanine dye has an absorbance of ca. 0.002 layer−1 at 532 nm, the interleaved films of 1b, reported in this study, are almost transparent and have a corresponding absorbance of ca. 5 × 10−4 bilayer−1 (Fig. 7).
![]() | ||
Fig. 7 Visible spectrum of 104 bilayers of the 1∶1 mixed film of 1b interleaved with poly(tert-butyl methacrylate). The residual absorbance at the higher wavelength end of the spectrum is attributed to the glass substrate. |
Mixed films of the dimethylamino and dibutylamino analogues show a similar quadratic SHG dependence to more than 100 bilayers. Their linear and nonlinear optical properties are summarised in Table 4 . The former has a lower second-order susceptibility of 40 pm V−1 (cf. 67 pm V−1 for 1b). This coincides with a reduced absorbance of 3 × 10−4 bilayer−1 at the harmonic wavelength and the absorption maximum being shifted to 439 nm (Fig. 8). Thus, resonant effects can explain the altered nonlinear optical properties and, although the spectra probably reflect slightly different packing arrangements, it is unclear why they are dependent upon the dialkylamino groups. Nonetheless, when interleaved with poly(tert-butyl methacrylate), dyes in this series form non-centrosymmetric structures. They exhibit SHG comparable with that obtained from conventional hemicyanine derivatives18–20 but have a greatly improved transmittance at the harmonic wavelength.
![]() | ||
Fig. 8 Visible spectrum of 105 bilayers of the 1∶1 mixed film of 1a interleaved with poly(tert-butyl methacrylate). |
Dye | Dye∶ODA | λmax/nm | φ/° | Film type | |
---|---|---|---|---|---|
1a | 1∶0 | 410 | 26 | 39 | interleaved |
1b | 1∶0 | 420 | 50 | 33 | interleaved |
1c | 1∶0 | 425 | 31 | 32 | Z-type |
1d | 1∶0 | 410 | 21 | 30 | Z-type |
1a | 1∶1 | 439 | 40 | 30 | interleaved |
1b | 1∶1 | 447 | 67 | 37 | interleaved |
1c | 1∶1 | 450 | 52 | 33 | interleaved |
1c | 1∶1 | 450 | 76 | 33 | Z-type |
![]() | ||
Fig. 9 Z-Type films of 1c: variation of the square root of the second-harmonic intensity with (a) the number of deposited layers and (b) the angle of incidence of the Nd∶YAG laser beam: p-polarised (solid circle) s-polarised (×) relative to the LB film. |
Another interesting feature of the 1∶1 mixed films of 1c and octadecanoic acid is that deposition on the downstroke (X-type) also results in non-centrosymmetric ordering but with the dipole orientation reversed. This has been demonstrated by X-type deposition onto a thick Z-type film, there being a suppression of the second-harmonic intensity when the number of layers of each type is equal. The intensity decreases quadratically with increasing film thickness and then, when the number of X-type layers is greater, it begins to increase quadratically. This is indicative of favourable LB deposition on the downstroke, as well as the upstroke, and may be used to fabricate waveguide structures with inversion symmetry in the thickness direction.16,28,29 The waveguiding properties of these films have not been studied but, in a parallel investigation on the 4-quinolinium analogue of 1c, we have previously reported Cerenkov-type phase matched SHG using fibre optic coupling.2,30 The materials reported here are more suitable candidates for investigation.
Footnote |
† Basis of a presentation given at Materials Chemistry Discussion No. 2, 13–15 September 1999, University of Nottingham, UK. |
This journal is © The Royal Society of Chemistry 2000 |